Reliability Enhancement of an Emergency Power Supply Benchmark based on Testing Optimization

This paper deals with testing interval optimization of two Diesel Generators (DGs) of an Emergency Power Supply System (EPSS) Benchmark. The proposed methodology is based on the fault tree development of the system and identifying their Minimal Cut Sets (MCS) using the professional RiskSpectrum PSA software. The failure on demand events of two DGs are modeled as a function of Testing Interval (TI). The Genetic Algorithms (GA) is adapted to optimize the TI of the DGs in order to reduce the failure probability of EPSS. The obtained results show a remarkable reduction of the failure probability of the system.